Antithrombotic therapy is recommended by guidelines for secondary prevention in ischemic stroke (IS). For patients with known cardioembolic IS, anticoagulation (AC) (e.g. with warfarin) is the antithrombotic therapy of choice. For all other IS including cryptogenic stroke (CS), antiplatelet therapy (APT) is preferred. We performed a recent meta-analysis including both randomized and (mostly) observational data from 8 studies suggesting a potential 50% relative risk reduction for AC versus APT in patients with cryptogenic stroke (CS) and patent foramen ovale (PFO). These results, however, are limited by low statistical power, confounding by indication and the lack of strong methods to control for non-comparability in the component observational studies. This proposal addresses these limitations by obtaining individual patient data (IPD) from the component studies and substantially augmenting the meta-analysis with data from our on-going Risk of Paradoxical Embolism (RoPE) Study (R01 NS062153). The RoPE Study focuses on the development of predictive models for the selection of patients that might benefit from PFO closure, including a model to estimate the probability that an index stroke is PFO-related. It includes 12 component databases of CS patients investigated for PFO. Seven of these RoPE database will be appropriate for this proposed TAcTiCS- PFO study. These 7 studies will be combined and harmonized with 3 to 5 additional non-RoPE databases, yielding a database with ~2000 patients. This is roughly 3-fold the number of patients included in our recent comprehensive literature-based study-level meta-analysis. More importantly, having IPD will permit us to use study-specific propensity models to adjust for confounding by indication. We hypothesize that, among patients with PFO and CS, AC may provide substantial benefit compared to APT, particularly in the subset of patients most likely to have stroke from paradoxical embolism, rather than occult atheroembolic or lacunar disease. Aim 1: To evaluate the comparative effectiveness of APT versus AC in patients with PFO and CS, using the largest combined database of medically-treated CS patients with PFO and strong methods to control for confounding by indication. Aim 2: To explore potential heterogeneity of treatment effect (HTE) in CS patients with PFO with a high versus a low probability of paradoxical embolism. Long term objectives: The results of this analysis are intended to motivate and inform a randomized clinical trial (RCT). Absent an RCT, they will provide the most definitive study available for selecting appropriate antithrombotic therapy in this important subgroup of IS patients. Public health significance: Stroke is the 3rd leading cause of death and leading cause of disability in the US. Among the ~795 000 people with stroke annually, ~185 000 are recurrent4. Secondary prevention is thus of enormous public health import. In most series, ~30% of strokes are classified as CS and about 50% of these have PFO, (i.e. >100,000 per year). Despite this, the optimal medical approach to secondary prevention in this population has not been well studied.